Quantum Physics II (8.05) Fall 2013
Assignment 4
Massachusetts Institute of Technology
Physics Department
September 27, 2013
Due October 4, 2013
3:00 pm
Problem Set 4
1. Identitites for commutators (Based on Griffiths Prob.3.13) [10 points]
In the following problem A, B, and C are linear operators. So are q and p.
(a) Prove the following commutator identity:
[A, BC] = [A, B] C + B [A, C] .
This is the derivation property of the commutator: the commutator with A,
that is the object [A, ·], acts like a derivative on the product BC. In the result
the commutator is first taken with B and then taken with C while the operator
that stays untouched is positioned at the expected place.
(b) Prove the Jacobi identity:
[[A, B] , C] + [[B, C] , A] + [[C, A] , B] = 0 .
(c) Using [q, p] = il and the result of (a), show that
[q n , p ] = il nq n−1 .
(d) For any function f (q) that can be expanded in a power series in q, use (c) to
show
[f (q), p] = ilf ′ (q) .
(e) On the space of position-dependent functions, the operator f (x) acts multiplica­
∂
tively and p acts as li ∂x
. Calculate [f (x), p ] by letting this operator act on an
arbitrary wavefunction.
2. Useful operator identities and translations [10 points]
Suppose that A and B are two operators that do not commute, [A, B] = 0.
(a) Let t be a formal variable. Show that
d t(A+B)
e
= (A + B) et(A+B) = et(A+B) (A + B) .
dt
1
Physics 8.05, Quantum Physics II, Fall 2013
2
(b) Now suppose [A, B] = c, where c is a c-number (a complex number times the
identity operator). Prove that
eA Be−A = B + c .
(1)
[Hint: Define an operator-valued function F (t) ≡ etA Be−tA . What is F (0)?
Derive a differential equation for F (t) and integrate it.]
Comment: Equation (1) is a special case of the Hadamard lemma, to be consid­
ered below.
(c) Let a be a real number and p̂ be the momentum operator. Show that the unitary
translation operator
p/l
T̂ (a) ≡ e−iaˆ
translates the position operator:
T̂ † (a) x̂ T̂ (a) = x̂ + a .
If a state |ψ) is described by the wave function (x|ψ) = ψ(x), show that the
state T̂ (a)|ψ) is described by the wave function ψ(x − a).
3. Proof of the Hadamard lemma [10 points]
Prove that for two operators A and B, we have
eA Be−A = B + [A, B] +
1
1
[A, [A, B]] + [A, [A, [A, B]]] + · · · .
2!
3!
(1)
Define f (t) ≡ etA Be−tA and calculate the first few derivatives of f (t) evaluated at
t = 0. Then use Taylor expansions. Calculating explicitly the first three derivatives
suffices to obtain (1).
Do things to all orders by finding the form of the (n + 1)-th term in the right-hand
side of (1). To write the answer in a neat form we define the operator ad A that acts
on operators X to give operators via the commutator
ad A(X) ≡ [A, X] .
Confirm that with this notation, the complete version of equation (1) becomes
eA Be−A = ead A (B) .
4. Special case of the Baker-Haussdorf Theorem [10 points]
Consider two operators A and B, such that [A, B] = cI, where c is a complex number
and I is the identity operator. You will prove here the following identity
eA+B = eB eA ec/2 = eA eB e−c/2 .
For this purpose consider the operator valued function
G(t) ≡ et(A+B) e−tA .
(2)
Physics 8.05, Quantum Physics II, Fall 2013
3
(a) Using operator properties and identities you derived previously show that
G−1
d
G(t) = B + ct .
dt
(3)
(b) Note that (3) is equivalent to dtd G(t) = G(t)(B + ct). Verify that the solution to
this equation is
1 2
G(t) = G(0) e tB e
2 ct .
(4)
(c) Consider G(1) to prove the first equality in (2). Rename the operators to obtain
the other equality.
Comment: The full Baker-Hausdorff formula is of the form
1
1
eX eY = e
X+Y + 2 [X,Y ]+ 12 ([X,[X,Y ]]−[Y,[X,Y ]])+...
and there is no simple closed form for general X and Y .
5. Bras and kets. [5 points]
Consider a three-dimensional Hilbert space with an orthonormal basis |1), |2), |3).
Using complex constants a and b define the kets
|ψ) = a|1) − b|2) + a|3) ;
|φ) = b|1) + a|2) .
(a) Write down (ψ| and (φ|. Calculate (φ|ψ) and (ψ|φ). Check that (φ|ψ) = (ψ|φ)∗ .
(b) Express |ψ) and |φ) as column vectors in the |1), |2), |3) basis and repeat (a).
(c) Let A = |φ)(ψ|. Find the 3 × 3 matrix that represents A in the given basis.
(d) Let Q = |ψ)(ψ| + |φ)(φ|. Is Q hermitian? Give a simple argument (no compu­
tation) to show that Q has a zero eigenvalue.
6. Shankar 1.8.8, p.43. Hermitian matrices and anticommutators [5 points]
7. Orthogonal projections and approximations (based on Axler) [15 points]
Consider a vector space V with an inner-product and a subspace U of V that is
spanned by rather simple vectors. (You can imagine this by taking V to be 3­
dimensional space, and U some plane going through the origin). The question is:
Given a vector v ∈ V that is not in U, what is the vector in U that best approxi­
mates v? As we also have a norm, we can ask a more precisely question: What is
the vector u ∈ U for which |v − u| is smallest. The answer is surprisingly simple: the
vector u is given by PU v, the orthogonal projection of v to U!
(a) Prove the above claim by showing that for any u ∈ U one has
|v − u| ≥ |v − PU v| .
Physics 8.05, Quantum Physics II, Fall 2013
4
As an application consider the infinite dimensional vector space of real functions in
the interval x ∈ [−π, π]. The inner product of two functions f and g on this interval
is taken to be:
� π
(f, g) =
f (x)g(x)dx .
−π
We will take U to be the a six-dimensional subspace of functions with a basis given
by (1, x, x2 , x3 , x4 , x5 ).
In this problem please use an algebraic manipulator that does integrals!
(b) Use the Gram-Schmidt algorithm to find an orthonormal basis (e1 , . . . , e6 ) for U.
(c) Consider approximating the functions sin x and cos x with the best possible
representatives from U. Calculate exactly these two representatives and write
them as polynomials in x with coefficients that depend on powers of π and
other constants. Also write the polynomials using numerical coefficients with
six significant digits.
(d) Do a plot for each of the functions (sin x and cos x) where you show the original
function, its best approximation in U calculated above, and the approximation
in U that corresponds to the truncated Taylor expansion about x = 0.
MIT OpenCourseWare
http://ocw.mit.edu
8.05 Quantum Physics II
Fall 2013
For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.